Printing of views without Netweaver 2004s (Solution 2)

Question

Hallo,
after writing my post concerning a printing solution before NetWeaver2004s (Solution 1),
i will also share my knowledge concerning my own printing solution developed inside Web Dynpro.
The following example should only demonstrate how an implementation of a print renderer can be done
inside Web Dynpro while you can take influence of the generated output.
I had to develop the possibility of printing some views including some head infos and table data.
The application is localized into six languages.
First i had a look at Adobe Document Services, but it is quite uncomfortable using table output
(it seems to be better supported in NetWeaver 2004s), but the problems resists in the localized labels.
You have to implement each lable also as an computed attribute to use it inside Adobe or
you have to copy the template for each language and translate it beside the xlf files.
I needed a solution that does not generate additional needs to print views.
So i decided to write an own PrintRendererClass using the Runtime view and the mapped context.
The following szenario should only demonstrate how you can archieve such a solution.
The Solution needs two Context attributes:
<b>ViewPrinter</b> (holds the reference to the renderer class)
<b>ViewPrinterURL</b> (String that receives an URL handle for using inside a LinkToURL control)
The instantiation of the class takes place inside the wdDoModifyView:
public static void wdDoModifyView(IPrivateDetailsView wdThis, IPrivateDetailsView.IContextNode wdContext,
     com.sap.tc.webdynpro.progmodel.api.IWDView view, boolean firstTime) {
//@@begin wdDoModifyView
if (firstTime) {
     ViewPrinter vprt = wdContext.currentContextElement().getViewPrinter();
     vprt = new ViewPrinter(view, wdContext, WDClientUser.getLoggedInClientUser().getLocale());
     vprt.setTitle(vprt.getTextById("Headline"));
     vprt.addGroup("GroupCustomerorder");
     vprt.addTable("TablePositions");
     wdContext.currentContextElement().setViewPrinterURL(vprt.render());                                           
//@@end
The RendererClass is only a framework for a big StringBuffer (xml) that uses methods to add
addtional output content to the string and finally generates a dynamic xhtml file.
The URL handle is returned at the end via vprt.render() and stored in the context attribute.
Inside the view you can use a LinkToURL control mapped to this context attribute,
which opens a new window with the generated xhtml file.
The most interessting parts are addGroup(uielementid) and addTable(uielementid).
The addGroup searches for a uielement with the given id and iterates through all entries.
For all found labels and their binded values output code will be added:
public void addGroup(String id) {
     // open group
     xml.append("<fieldset>");
     IWDGroup group = (IWDGroup) view.getElement(id);
     // get group header text
     String legend = group.getHeader().getText();
     xml.append("<legend>" + legend + "</legend>");
     // trace all elements
     for (Iterator iter = group.iterateChildren(); iter.hasNext();) {
          IWDUIElement uiEl = (IWDUIElement) iter.next();
          if (uiEl instanceof IWDLabel) {
               xml.append("<div>");
               this.addLabelWithValue(uiEl.getId());
               xml.append("</div>");                    
     // close group
     xml.append("</fieldset>");
For each found label the label and binded text should be printed out:
public void addLabelWithValue(String id) {
     // find given label (with value)
     IWDLabel lbl = (IWDLabel) view.getElement(id);
     if (lbl!=null) {
          xml.append("<label for="" + lbl.getLabelFor() + "">" +
               encodeHTML(lbl.getText(),true) + ": </label>");
          // find corresponding value          
          IWDTextView txt = (IWDTextView) view.getElement(lbl.getLabelFor());
          if (txt!=null) {
               xml.append("<span id="" + lbl.getLabelFor() + "">" +
                    encodeHTML(txt.getText(),true) + "</span>");
The most interesting part is the generic output of a table:
public void addTable(String id) {
     // find given table
     IWDTable table = (IWDTable) view.getElement(id);
     String tableBindedDataSource = table.bindingOfDataSource(); // node "Positions"
     xml.append("<fieldset>")
          .append("<legend>" + table.getHeader().getText() + "</legend>");
     xml.append("<table summary="")
          .append(table.getHeader().getText())
          .append("">");
     xml.append("<thead><tr>");
     // use iterator          
     ArrayList cellAlign = new ArrayList();     
     ArrayList cellId = new ArrayList();
     int nrUsedColumns = 0;
     for (Iterator iter = table.iterateColumns(); iter.hasNext();) {
          IWDTableColumn col = (IWDTableColumn) iter.next();
          // check column id against blacklist
          if (!blackList.contains(col.getId())) {                    
               IWDTableCellEditor cellEditor = col.getTableCellEditor();
               // casting for supported uiwidgets
               String bindedColumnText = null;
               if (cellEditor instanceof IWDTextView) {
                    IWDTextView txt = (IWDTextView) cellEditor;               
                    bindedColumnText = txt.bindingOfText(); // "TABLE.ATTRIBUTE"                                                  
               // add supported widgets to print
               if (bindedColumnText != null) {
                    // remove leading TABLE prefix (use only ATTRIBUTE)               
                    String textId = bindedColumnText.substring(bindedColumnText.lastIndexOf('.')+1);
                    // get column alignment
                    String colAlign = col.getHAlign().toString().toLowerCase();
                    String classAlign = (colAlign.equalsIgnoreCase("auto"))
                         ? "" : " class="" + colAlign + """;
                    xml.append("<th" + classAlign + " scope="col" abbr="" + col.getHeader().getText() + "">")
                         .append(col.getHeader().getText())
                         .append("</th>");
                    // make column values for table body available
                    cellAlign.add(nrUsedColumns, colAlign);
                    cellId.add(nrUsedColumns, textId);
                    nrUsedColumns++;
     xml.append("</tr></thead>");
     // print out table body
     xml.append("<tbody>");          
     // table corresponding context node
     IWDNode node = ((IWDNode)wdContext).getContext().getRootNode().
          getCurrentElement().node().getChildNode(tableBindedDataSource, 0);
     int tableSize = node.size();     
     IWDNodeElement nodeEl = null;
     for (int i = 0; i < tableSize; i++) {
          xml.append("<tr>");     
          nodeEl = node.getElementAt(i);
          for (int col = 0; col < nrUsedColumns; col++) {     
               Object value = nodeEl.getAttributeValue(cellId.get(col).toString());                    
               String classAlign = (cellAlign.get(col).toString().equalsIgnoreCase("auto"))
                    ? "" : " class="" + cellAlign.get(col).toString() + """;               
               xml.append("<td" + classAlign + ">"
                    + encodeHTML(this.getAttributeValue(value), true) + "</td>");
          xml.append("</tr>");
     xml.append("</tbody>");
     xml.append("</table>");
     xml.append("</fieldset>");     
[code]
First we inspect all table columns to get the head labels and the used cell editors.
There is also support for a black list to allow skipping of columns that should not be printed.
Invisible columns are also printed, so you can add addtional table columns for printing that are
disabled for the browser. Also the alignment of the cell editor will be supported.
A second loop iterates over the table context depending the used columns.
To get the formatted values concerning their datatype i use the method getAttributeValue,
cause i think the formatting of mapped Context to the browser takes place in the deepest parts
of the Web Dynpro Renderer Task and is not accessible via API.
Therefore you have to implement formatters for all supported datatypes:
[code]
private String getAttributeValue(Object obj) {          
     String txt = "";
     if (obj != null) {     
          // String handling
          if (obj instanceof java.lang.String) {
               txt = obj.toString();
          // Date handling     
          } else if (obj instanceof java.sql.Date) {
               DateFormat formater = DateFormat.getDateInstance(
                    DateFormat.SHORT, this.locale);
               txt = formater.format(obj);
          // BigDecimal handling
          } else if (obj instanceof java.math.BigDecimal) {     
               DecimalFormat df =
                    (DecimalFormat)DecimalFormat.getInstance(this.locale);
               df.applyPattern( "0.######" );
               txt = df.format(obj);
          // Default Handling -> with debug output of Class Name
          } else {
               txt = "<span class="pointer red" title="" + obj.getClass().getName() +
                    "">" + obj.toString() + "</span>";
     return txt;
Here is the reason, why the class constructor needs a locale!
I try to format the output concerning the portal user language.
At the end you have to call the vprt.render() method to start creation of a new html ressource.
The method returns the ressource url as a string which is binded to a LinkToUrl control.
That's it.
The above solution only fullfills my needs concerning the project. I implemented only the
needed methods to print out groups and tables.
The table renderer skips all table columns that are not IWDTextView.
if you need addtional things like IWDProgressBar, you have to implement such a kind of renderer.
The solution is a nice way to generate printable content of views with the following benefits
Pro:
- Print out is Locale/Language dependend (Text, Date, ...)
- You can influence the print design via css
- You can implement your own controll renderer (like decorators)
- It works for SAP Net Weaver releases <= 2004s
Cons:
- To support all widgets, you have to implement them
  (controls/widgets are growing with new service packs)
The functionallity shows how you can use the abstract api for iterating
through view elements and their mapped context in a generic way.
The solution is a plain structured class because i do not want to reimplement
the whole Widgets API for printing a view (not in the scope/budget of our customers).
If you are interessted in the working class file, please contact me.
Maybe this helps someone who has to do similarly things inside Web Dynpro
and does not want to reinvent the wheel.
Best wishes,
Holger Schaefer

Answer 1

Hi Holger,
This Solution helped me a lot. Thanks a lot.
Sreekanth